Jetson NX development board reboots randomly

Hello

I have noticed my Jetson NX development board reboots when I run applications. Serial connection does not show any message before these random reboots. I have not connected any peripherals to the board.

I appreciate any help to resolve this issue. Thank you

1 Like

hello kalindu.herat,

may I know which JetPack release you’re used, also, what’s the application or use-case you’re running with.
thanks

Hi Jerry Chang

Thank you very much for response
Running “cat /etc/nv_tegra_release” has given me the following output.

R32 (release), REVISION: 4.3, GCID: 21589087, BOARD: t186ref, EABI: aarch64, DATE: Fri Jun 26 04:34:27 UTC 2020

I usually run python (3.6) applications. I also access Performance Monitoring Unit (PMU) of the CPU using Linux Perf. Python application processes the PMU data I obtain using Perf.

Again, Thank you very much

hello kalindu.herat,

please execute tegrastats utility with sudo permission to gather the processor usage statistics for reference.
thanks

Dear Jerry Chang

Thank you for the response. This is what I got from the tegrastats utility just before the reboot.

RAM 622/7772MB (lfb 1470x4MB) SWAP 0/3886MB (cached 0MB) CPU [2%@1190,0%@1190,off,off,off,off] EMC_FREQ 0%@204 GR3D_FREQ 0%@114 APE 150 MTS fg 0% bg 0% AO@29.5C GPU@30C PMIC@100C AUX@29C CPU@30.5C thermal@29.75C VDD_IN 1628/2772 VDD_CPU_GPU_CV 122/617 VDD_SOC 366/744

hello kalindu.herat,

okay, this issue is not caused by throttling, since it shows very light usage.
may I know what’s the power-supply you’re using, you should have 19V power supply according to Jetson Xavier NX Developer Kit | NVIDIA Developer,
thanks

Dear Jerry Chang

The power I use is the original adapter that comes with Jetson NX Development kit. The rating of the adapter is 19.0 V, 2.37 A.

Thanks a lot.

hello kalindu.herat,

you may review the process. for example, please check you’re access an invalid address?
thanks

Dear Jerry Chang

It seems the reboot occurs when the application accesses an invalid address. Thank you very much for directing me. I have marked your response as the solution.

The reboot due to accessing invalid access is not immediate. May I know is this reboot occurs due to a hardware routine or a OS routine?

Thanks again.

hello kalindu.herat,

it should be a segmentation fault for an illegal access, you may using $ dmesg --follow for checking the details.
thanks

Hello Jerry Chang

I observed dmesg and observed that “No way out” has stopped resetting Watchdog Timer. This has rebooted the device after awhile.
To overcome this situation, I have rebuilt the kernel with CONFIG_WATCHDOG_NOWAYOUT=0.

Now the device does not reboot.

Thank you very much for your guidance.